Towards bi-directional dancing interaction

Dancing is an entertaining form of taskless interaction. When interacting with a dancing Embodied Conversational Agent (ECA), the lack of a clear task presents the challenge of eliciting an interaction between user and ECA in a different way. In this paper we describe our Virtual Dancer, which is an ECA that invites a user to dance. In our system the user is monitored using global movement characteristics from a camera and a dance pad. The characteristics are used to select and adapt movements for the Virtual Dancer. This way, the user can dance together with the Virtual Dancer. Any interaction patterns and implicit relations between the dance behaviour of the human and the Virtual Dancer should be evoked intuitively without explicit appeal. The work described in this paper can be used as a platform for research into natural animation and user invitation behavior. We discuss future work on both topics.